Abstract
A method for analysing matched case control data in presence of overdispersion or correlation between observations within matched sets is presented. A generalised linear mixed model approach is adapted to account for overdispersion by allowing random exposure effects. Expressions for the elements of the estimation equation are derived and a strategy for achieving efficiency in computing is discussed. Results from simulation studies and two applications are presented.
Original language | English |
---|---|
Pages (from-to) | 271-281 |
Number of pages | 11 |
Journal | Biometrical Journal |
Volume | 43 |
Issue number | 3 |
DOIs | |
Publication status | Published - 2001 |